bioRxiv preprint doi: https://doi.org/10.1101/781765; this version posted September 29, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. PHD3 controls energy homeostasis and exercise capacity Haejin Yoon1, Jessica B. Spinelli1, Elma Zaganjor1, Samantha J. Wong1, Natalie J. German1, Elizabeth C. Randall 2, Afsah Dean3, Allen Clermont3, Joao A. Paulo1, Daniel Garcia4, Hao Li5, Nathalie Y. R. Agar2, 6, 7, Laurie J. Goodyear3, Reuben J. Shaw4, Steven P. Gygi1, Johan Auwerx5, Marcia C. Haigis1* 1Department of Cell Biology, Harvard Medical School, Boston, 2Department of Radiology, Brigham and Women’s Hospital, Harvard Medical School, Boston, Boston, 3Section on Integrative Physiology and Metabolism, Joslin Diabetes Center, Boston, 4The Salk Institute for Biological Studies, 10010 N. Torrey Pines Road, La Jolla, 5Laboratory of Integrative and Systems Physiology, École Polytechnique Fédérale de Lausanne, Lausanne 1015, Switzerland, 6Department of Neurosurgery, Brigham and Women’s Hospital, Harvard Medical School, Boston, Department of Cancer Biology, 7Dana-Farber Cancer Institute, Harvard Medical School, Boston. *Correspondence to [email protected] SUMMARY Rapid alterations in cellular metabolism allow tissues to maintain homeostasis during changes in energy availability. Acetyl-CoA carboxylase 2 (ACC2) provides a central regulation and is phosphorylated acutely by AMPK during cellular stress to relieve repression of fat oxidation. While ACC2 is hydroxlated by prolyl hydroxylase 3 (PHD3), the physiological consequence of PHD3 is little understood. We find that ACC2 phosphorylation and hydroxylation occur in a reciprocal fashion. ACC2 hydroxylation occurs in conditions of high energy and leads to decreased fatty acid oxidation. Furthermore, AMPK-mediated phosphorylation of ACC2 is bioRxiv preprint doi: https://doi.org/10.1101/781765; this version posted September 29, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. inhibitory to hydroxylation. PHD3 null mice demonstrate loss of ACC2 hydroxylation in heart and skeletal muscle and display elevated fatty acid oxidation. Whole body or skeletal muscle- specific PHD3 loss enhanced exercise capacity during an endurance exercise challenge. In sum, these data identify PHD3 as a physiological regulator of energy balance during acute stress. INTRODUCTION Metabolic adaptation plays a fundamental role in maintaining energy homeostasis. In many tissues, fatty acids are exploited as an adaptive fuel to enable survival under conditions of metabolic stress, such as nutrient deprivation or exercise (Galgani and Ravussin, 2008; Woods and Ramsay, 2011; Palm and Thompson, 2017; Efeyan et al., 2015). As one example of metabolic adaptation, cells utilize fatty acid oxidation (FAO) to respond to low energy conditions, cold temperature, oxidative stress, and exercise (Lee et al., 2016; Chouchani et al., 2016; Grunt, 2018; Herzig and Shaw, 2018). However, the mechanism for the dynamic regulation of fatty acid oxidation in response to metabolic stress is not completely understood. A reduction in cellular ATP and increase in the AMP/ATP ratio leads to activation of AMP-activated protein kinase (AMPK) (Hardie et al., 2012; Gowans and Hardie, 2014). In response to high AMP/ATP, AMPK phosphorylates acetyl-CoA carborylase (ACC) 1/2. AMPK phosphorylation of ACC 1/2 inhibits its activity to convert acetyl-CoA into malonyl-CoA. In the cytosol, ACC1 generates pools of malonyl-CoA thought to be important for de novo lipogenesis, while ACC2 associated with the outer mitochondrial membrane generates malonyl-CoA to inhibit carnitine palmitoyltransferase 1 (CPT1), which mediates the first step of long chain fatty acid transport into the mitochondria. Thus, ACC2 phosphorylation by AMPK provides a readout of cellular energy/stress with direct ties to fat synthesis and oxidation. ACC2 is also subject to positive regulation in the control of fatty acid metabolism via hydroxylation on P450 by proline hydroxylase domain protein 3 (PHD3) (German et al., 2016). Prolyl hydroxylase domain family members (PHDs, also called EGLN1-3) are alpha- ketoglutarate-dependent dioxygenases (Epstein et al., 2001). A previous study showed that tumors with low PHD3 exhibited decreased ACC2 hydroxylation and elevated fatty acid oxidation (German et al., 2016). By contrast, the role of PHD3 in the control of fat oxidation in energy homeostasis has not been examined fully. bioRxiv preprint doi: https://doi.org/10.1101/781765; this version posted September 29, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Here, we show that PHD3 modifies ACC2 at proline 450 physiologically in response to nutrient and energy availability in cells and mouse tissues, and impacts lipid and energy homeostasis in whole body PHD3 KO mice and skeletal muscle-specific PHD3 KO mice. Intriguingly, AMPK-mediated phosphorylation of ACC2 inhibits PHD3 binding and hydroxylation of ACC2, demonstrating one mechanism for the reciprocal relationship between ACC2 hydroxylation and phosphorylation. Using whole body or muscle specific PHD3 null mice, we probe the physiological relevance of PHD3 during fasting and exercise challenges. PHD3 loss increases fatty acid oxidation and the loss of PHD3 is sufficient to increase exercise capacity. Together, these results suggest that PHD3 signaling may have potential to tune fatty acid oxidation in response to energy state. RESULTS ACC2 hydroxylation is sensitive to glucose and negatively regulated by AMPK In this study we sought to examine a physiological role for PHD3 in the control of fatty acid oxidation. We previously reported that PHD3 physically interacts with ACC2, a central regulator of fat catabolism (German et al, 2016). While both PHD3 and ACC2 signaling potentially converge on ACC2 through distinct post-translational modifications (PTMs), whether these PTMs act in a concerted manner to regulate ACC2 activity has not been explored (Figure 1A). As PHD3 is an alpha-ketoglutarate- dependent dioxygenase, and activation of ACC2 would subsequently repress FAO, we reasoned that PHD3 may hydroxylate ACC2 in response to nutrient availability. To examine whether PHD3 activity was sensitive to fluctuations in nutrient availability, we probed ACC2 hydroxylation in mouse embryonic fibroblasts (MEFs) cultured in the presence of low (5 mM glucose) or high glucose (25 mM glucose) in the presence of 10% FBS (Figure 1B). Since AMPK is active by increased AMP/ATP conditions, we utilized low glucose culture media as a control (Salt et al., 1998; Laderoute et al., 2006). As expected, ACC2 phosphorylation was increased in the low glucose condition. By contrast, we observed a reciprocal hydroxylation in ACC2 (Figure 1B and 1C) as detected by reactivity with an antibody to proline hydroxylation. Importantly, the addition of 25 mM glucose caused ACC2 hydroxylation within 5 minutes, but not in cells treated with the pan-PHD inhibitor dimethyloxallyl glycine (DMOG) (Figure 1B and 1C). DMOG treatment significantly decreased bioRxiv preprint doi: https://doi.org/10.1101/781765; this version posted September 29, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. hydroxyl ACC2 levels compared to untreated control cells in high glucose media (Figure 1B and 1C). Finally, we tested whether other energetic inputs affected ACC2 hydroxylation (Figure bioRxiv preprint doi: https://doi.org/10.1101/781765; this version posted September 29, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. S1A). In contrast to glucose, the addition of amino acids, fatty acids or dialyzed serum resulted in little to no induction of ACC2 hydroxylation (Figure 1D, Figure S1B-C). In sum, these data suggest a rapid induction in ACC2 hydroxylation in response to glucose abundance. We utilized Tandem Mass Tag (TMT) mass spectrometry to quantify changes in ACC2 proline modifications in response to glucose or PHD3 (Paulo, 2016; Navarrete-Perea et al., 2018; Berberich et al., 2018). In ACC2 immunoprecipitates, we observed a number of proline residues modified with a +15.9949 Th shift (Figure S1D), consistent with modification by hydroxylation. However, only P450 demonstrated elevated signal of a +15.9949 Th shift in a glucose- or PHD3- sensitive manner. The P450 +15.9949 Th shift was diminished 5.3-fold in PHD3 KD cells compared with control cells, and the P450 15.9949 Th shift was reduced 2.3-fold in wildtype control cells cultured in low glucose compared with high glucose media (Figure 1E, Figure S1E- G). These data are similar in magnitude to the fold-changes observed in ACC2 hydroxylation detected by Western blotting (Figure S1H). Of interest, our studies suggested that ACC2 was phosphorylated in a reciprocal manner to its proline modification (Figure 1D). This observation suggested some level of circuitry between AMPK and PHD3 activity, and so we aimed to understand whether AMPK influenced PHD3 activity or vise versa. First, we tested whether PHD3 activity was affected by AMPK loss of function by measuring